76 



BULLETIN 60, UNITED STATES NATIONAL MUSEUM. 



The following arrangement will illustrate the progression of special- 

 ization in five of the principal characters in the order: 



Name. 



Mengea 



Trioxocera . 



Myrmecolax. 

 Caenocholax. . 



Stylops 



Hylechthrus. 



Halictoxenos . 



Crawfnrdia. 



Dioxocera. 



Halictophagus. . 



Pentoxocera 



A nthericommr . 



Mecynocera. 

 Elenchus 



Elenchoides . 



Males. 



Number of tarsal 

 joints. 



Xenos 



Apractelytra 



Pseudoxenos 



Monobiaphila | 



Leionotoxenos j 



Paraxenos ;J 



Eupathocera | ( 



Ophthalmochlus. . . 



Sehistosiphon 



A croschismus 



Homilops 



Mengcnidea. 

 5 

 5 

 Xenoidea. 

 4 

 4 



4 



4 



Ilalictophagnidea . 

 3 



3 

 3 

 3 

 Elenchoidea. 

 2 

 2 

 (?) 



Number of antennal 

 joints. 



Mengeidse. 



7 



MyrmecolocidsR. 



Stylopidse. 



6 



Ilylechthridae. 



5 



Xenidx 

 4 



4 

 4 



Dioxoceridse . 

 4 

 Halictophagidx. 

 7 

 7 

 7 

 Elenchidse. 

 5 

 5 

 (?) 



Number of 



primary veins 



in wings. 



(?) 



8 

 8 



(?) 



. (?) 



Crawfordinse. 



5 



(?) 



6 

 6 

 5 



4 



(?) 

 (?) 



Females. 



Number of 

 genital 

 tubes. 



(?) 

 (?) 



(?) 

 (?) 



(?) 



Halictox- 



aiinx. 



5 

 Xeninx. 

 4 



(?) 

 (?) 



(?) 



(?) 

 (?) 



4 

 (?) 



(?) 



(?) 



(?) 

 (?) 

 (?) 



(?) 

 (?) 

 3 



Trlun- 

 gulinids. 



Number 

 of simple 

 segments 

 in ab- 

 domen. 



(?) 

 (?) 



(?) 

 (?) 



(?) 



9 



8 



(?) 

 8 



(?) 



8 



(?) 

 8 

 8 



(?) 



(?) 



(?) 

 8 

 (?) 



(?) 



(?) 



The two points which stand out in this comparison are the affinities 

 of Halictoxenos to the Stylopidse as well as Xenidse, and the interme- 

 diate position of Dioxocera between the Xenidae and Halictophagidse. 

 The grouping of genera in the Xenidse is very weak at present, owing 

 to the deficiencies in our knowledge, hence the best possible way of 

 grouping the genera at present is by host relationships. Many sur- 

 prises are still in store in this order, and the classification will un- 

 doubtedly need many revisions. The group Homilopinae rests upon 

 a description which may prove wrong; in such event Homilops will 



